Diaphragm-based pressure sensors have been used for a variety of applications, where pressure exerted by a pressurized medium deflects a diaphragm, and sensing elements (such as strain gauges) coupled to the diaphragm sense the deflection and provide a signal correlating the deflection of the diaphragm with the amount of pressure.
There are two major types of pressure sensors. The first type is called a gauge pressure sensor, which measures pressure with respect to atmospheric pressure. The second type is called an absolute pressure sensor, which typically measures pressure with respect to a vacuum or zero pressure.
FIGS. 1A and 1B respectively show a conventional gauge pressure sensor 100 and a conventional absolute pressure sensor 110, both by Silicon Microstructures, Inc. of Milpitas, Calif. (pressure sensor models SM5102). Both the pressure sensors 100 and 110 use a silicon micromachined structure (also known as a gauge wafer) with a diaphragm having sensing elements (not shown) on the top or outer surface. The micromachined structure is mounted on a support structure (also known as a spacer). Typical dimensions of the pressure sensors are shown in millimeters. Both the pressure sensors 100 and 110 are top side pressure sensors, i.e., the top or outer side of the diaphragm is accessed by the pressurized medium. The support structure in the gauge pressure sensor 100 has an opening to expose the opposite side (i.e., the bottom side or the inner side) of the diaphragm to atmospheric pressure. On the other hand, the support structure in the absolute pressure sensor 110 has no opening, and defines a vacuum reference cavity underneath the diaphragm. Both the pressure sensors 100 and 110 are not ideal for applications involving harsh pressure media (such as fuel mixtures, acidic solution, and the like), as the sensing elements on the top side of the diaphragm may come in contact with the harsh pressurized medium if a protective coating on the sensing elements is damaged.
The gauge pressure sensor reads pressure with respect to atmospheric pressure. Atmospheric pressure varies over elevation and weather conditions. Thus, an absolute pressure sensor is often preferred where high accuracy is needed. For example, gauge pressure reading can change by 2-3 psi due to variations in atmospheric pressure. Thus, it can contribute to 2% to 3% error in the pressure reading if the full scale is 100 psi. Most new pressure sensor applications below 500 psi require +/−1% accuracy over operational temperature range, pressure range, and over the life of the product. Thus absolute pressure sensing is becoming very important in such applications.
FIG. 2 shows a conventional absolute pressure sensor 200, by Silicon Microstructures, Inc. of Milpitas, Calif. (pressure sensor model SM5112). Pressure sensor 200 uses a silicon gauge wafer with a micromachined diaphragm. A boro-silica glass cap sealed to the gauge wafer creates a reference vacuum cavity on top of the diaphragm. The pressurized medium exerts pressure from the bottom side of the diaphragm. The support structure (spacer) is made of boro-silica glass with a drilled hole in the center to allow the pressurized medium to access the bottom side of the diaphragm. The top side of the diaphragm has strain gauge sensing elements, interconnecting diffused resistors, and electrical interconnect metallizations to bring electrical signals out from the strain gauge sensing elements. This configuration is better for applications involving harsh pressurized media, as the sensing elements are separated from the pressurized medium. However, attaching a boro-silica glass cap wafer to a silicon gauge wafer to create a vacuum cavity is a very expensive process.
Another problem encountered by pressure sensor 200 is failure to withstand high temperature because of electrical leakage in the sensing elements. Pressure sensor 200 uses piezoresistors as sensing elements (configured as a strain gauge). In a typical piezoresistive strain gauge, four piezoresistors are connected in a Wheatstone bridge configuration (see FIG. 4) on top of the diaphragm. When the applied pressure deflects the diaphragm, induced stress in the diaphragm causes the piezoresistors to change their respective resistance values, resulting in an imbalance in the Wheatstone bridge. The imbalanced piezoresistor bridge produces an electrical signal output that is proportional to the applied pressure. In a silicon diaphragm, piezoresistors can be integrated at a low cost by using standard photolithographic processes. As shown in FIG. 3 (only the gauge wafer is shown here), piezoresistors may be defined as diffused wells of opposite-polarity regions embedded in the bulk material of the diaphragm. In the inset of FIG. 3, p-type diffused piezoresistors are created in the n-type bulk silicon diaphragm by photolithographically opening windows in the top insulator layer, and then doping with p-type material (e.g., boron) to achieve a desired sheet resistivity. The piezoresistors create diode-like p-n junctions with the diaphragm substrate, as shown in the equivalent circuit in FIG. 4. At temperatures beyond about 125° C., the p-n junctions behave like leaky diodes with the leakage current increasing exponentially as the temperature rises. As shown in FIG. 4, if the surrounding pressurized medium is electrically conductive, current leaks to the ground through the pressurized medium (represented by the resistor R in the equivalent circuit). The current leakage is substantial at higher temperatures, and causes sensing malfunction and may cause irreversible physical damage. Although not shown in FIG. 3 for the sake of clarity, persons of ordinary skill in the art will now understand that interconnecting diffused resistors (as shown in FIG. 2) coupled to the strain gauge piezoresistors will also contribute to current leakage at high temperature. Moreover, contamination during fabrication processes may cause diode current leakage even at lower temperatures such as room temperature.
In the previously discussed examples, a silicon diaphragm is used with integrated sensing elements. Silicon diaphragms and integrated sensing elements are popular because of ease of manufacturing using batch processing. However, depending on particular applications, in some conventional pressure sensors, the diaphragm and the sensor may be separated. This may be useful from a harsh pressurized media compatibility standpoint, as the diaphragm can be made of a corrosion-resistant material, such as stainless steel, and the sensing element can be made of silicon and can be kept isolated from pressurized medium exposure in a sealed chamber filled with an additional pressure transfer medium. One example of this type of sensor is referred to as an oil-filled sensor, where the pressure transfer medium is oil. This process can be relatively expensive, as the oil filling has to be performed in a vacuum. Errors arise in this approach because there is usually a small amount of residual air in the chamber after sealing. Thermal effects on the oil volume and air bubble also act to increase the error in the pressure reading. For reference, readers are encouraged to read U.S. Pat. No. 6,311,561 to Bang et. al.
In the case of a corrosion-resistant metal diaphragm, strain gauges are usually defined by depositing and patterning thin metal films on the diaphragm. For example, a titanium oxy nitride (TiON) strain gauge layer may be deposited on a silicon dioxide coated stainless steel diaphragm. However, typically, this type of strain gauge has lower gauge factors than micromachined silicon piezoresistive strain gauges, affecting pressure measurement accuracy.
By adopting a hybrid configuration, where a micromachined silicon piezoresistive strain gauge is bonded to an oxide-coated metal diaphragm, one can address the low gauge factor problem. However, the hybrid pressure sensor suffers from a thermal expansion mismatch problem between the sensing elements and the diaphragm. Moreover, the hybrid construction may not be efficient for batch processing. Adding a vacuum sealed cavity on top of the sensing elements can be prohibitively expensive. Even depositing oxide on the stainless-steel diaphragm requires an expensive fine polishing process. The operating temperature is typically limited to 140° C. in the hybrid pressure sensor.
Absolute pressure sensors, where the sensing elements are enclosed in a sealed reference cavity, offer the advantage of protection of sensing elements from harsh pressurized media. However, special design and processing steps are required to bring out electrical connections from the sensing elements to outside the sealed chamber, i.e. packaging of the sensor becomes costly. U.S. Pat. Nos. 5,929,497 and 6,109,113 show one way of bringing out electrical connections from a vacuum cavity. The process is complicated and uses capacitive sensors and poly-silicon connections. Wafers are bonded using an electrostatic bonding technique. The same technique with additional circuits is described in U.S. Pat. No. 6,713,828. These techniques are suited for ambient and sub-atmospheric pressure levels in motor vehicle applications.
Accordingly, a new pressure sensor, that offers a combination of a plurality of desired features, including, but not limited to, high accuracy absolute pressure sensing, wide pressure range, chemical and electrical compatibility with harsh pressurized media, reliable operation at all ranges of temperature including high temperatures, ease of manufacturing and packaging, compact size, and low cost would be desirable.